qlora-adapter-Llama-2-7b-hf-databricks-dolly-15k

This model is a fine-tuned version of meta-llama/Llama-2-7b-hf on the databricks/databricks-dolly-15k dataset.

It achieves the following results on the evaluation set:

• Loss: 1.1313

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Trained on RTX A5000 - 24GB GPU. The training took 3 hours 31 mins on the datasets with 12008 train samples and 1501 validation samples

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0001
• train_batch_size: 1
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss
1.1584	0.08	1000	1.1782
1.0667	0.17	2000	1.1710
1.0662	0.25	3000	1.1599
1.0517	0.33	4000	1.1569
1.0479	0.42	5000	1.1502
1.0516	0.5	6000	1.1441
1.0612	0.58	7000	1.1397
1.0235	0.67	8000	1.1361
1.0259	0.75	9000	1.1339
1.0485	0.83	10000	1.1320
1.0406	0.92	11000	1.1314
1.0393	1.0	12000	1.1313

Framework versions

• Transformers 4.33.3
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.13.3